updating to local archived copy of openfst and Debian Bookworm

.github/workflows/CI.yml

@@ -8,8 +8,12 @@ jobs:
   # First we build the docker container, then we run CI tests
   build:
     runs-on: ubuntu-latest
+    strategy:
+      matrix:
+        debian_base: [bullseye, bookworm]
     env: 
       DOCKER_IMAGE: fstalign
+      DEBIAN_BASE: ${{ matrix.debian_base }}
 
     steps:
       - name: Checkout repository and submodules
@@ -18,7 +22,10 @@ jobs:
           submodules: recursive
 
       - name: Build the docker container
-        run: docker build . -f Dockerfile -t ${DOCKER_IMAGE}
+        run: |
+          docker build . -f Dockerfile \
+            --build-arg DEBIAN_BASE=${DEBIAN_BASE} \
+            -t ${DOCKER_IMAGE}:${DEBIAN_BASE}
 
       - name: Run CI tests
-        run: docker run --rm -t ${DOCKER_IMAGE} bash -c '(cd build && make test)'
+        run: docker run --rm -t ${DOCKER_IMAGE}:${DEBIAN_BASE} bash -c '(cd build && make test)'

.github/workflows/deploy.yml

@@ -6,6 +6,11 @@ jobs:
   push_to_registry:
     name: Push Docker image to GitHub Packages
     runs-on: ubuntu-latest
+    strategy:
+      matrix:
+        debian_base: [bullseye, bookworm]
+    env:
+      DEBIAN_BASE: ${{ matrix.debian_base }}
     steps:
       - 
         name: Check out the repo
@@ -22,7 +27,10 @@ jobs:
           if [[ $GITHUB_REF == refs/tags/* ]]; then
             VERSION=${GITHUB_REF#refs/tags/}
           fi
-          TAGS="${DOCKER_IMAGE}:${VERSION}"
+          TAGS="${DOCKER_IMAGE}:${VERSION}-${DEBIAN_BASE}"
+          if [ "${DEBIAN_BASE}" = "bullseye" ]; then
+            TAGS="${TAGS} ${DOCKER_IMAGE}:${VERSION} ${DOCKER_IMAGE}:latest"
+          fi
           echo ::set-output name=version::${VERSION}
           echo ::set-output name=tags::${TAGS}
           echo ::set-output name=created::$(date -u +'%Y-%m-%dT%H:%M:%SZ')
@@ -45,6 +53,6 @@ jobs:
           context: .
           platforms: linux/amd64
           push: true
-          tags: |
-            revdotcom/fstalign:latest
-            ${{ steps.prep.outputs.tags }}
+          build-args: |
+            DEBIAN_BASE=${{ env.DEBIAN_BASE }}
+          tags: ${{ steps.prep.outputs.tags }}

Dockerfile

@@ -1,19 +1,54 @@
-# Using kaldi image for pre-built OpenFST, version is 1.7.2
-FROM kaldiasr/kaldi:cpu-debian10-2024-07-29 as kaldi-base
+ARG DEBIAN_BASE=bullseye
+FROM debian:${DEBIAN_BASE}-slim AS debian-base
+RUN echo "APT::Get::Assume-Yes \"true\";\nAPT::Get::allow \"true\";" | tee -a  /etc/apt/apt.conf.d/90_no_prompt && \
+    echo "APT::Keep-Downloaded-Packages \"false\";" | tee -a  /etc/apt/apt.conf.d/91_no_cache && \
+    apt-get update
+# Stage 1: Build OpenFST 1.7.2 from source
+FROM debian-base as openfst-builder
 
-FROM debian:11
+ARG OPENFST_VERSION=1.7.2
+ARG JOBS=4
+
+# Install build dependencies for OpenFST
+RUN apt-get update && \
+    apt-get upgrade -y && \
+    apt-get install -y --no-install-recommends \
+    g++ \
+    make \
+    ccache \
+    && rm -rf /var/lib/apt/lists/*
+
+# Copy and build OpenFST from local tarball
+WORKDIR /tmp
+COPY ext/openfst-${OPENFST_VERSION}.tar.gz /tmp/
+RUN --mount=type=cache,target=/root/.ccache,sharing=locked  \
+    PATH=/usr/lib/ccache:${PATH} \
+    tar -xzf openfst-${OPENFST_VERSION}.tar.gz && \
+    cd openfst-${OPENFST_VERSION} && \
+    ./configure --prefix=/opt/openfst --enable-shared --enable-static && \
+    make -j${JOBS} && \
+    make install && \
+    cd .. && \
+    rm -rf openfst-${OPENFST_VERSION} openfst-${OPENFST_VERSION}.tar.gz
+
+# Stage 2: Build fstalign
+FROM debian-base
 
-COPY --from=kaldi-base /opt/kaldi/tools/openfst /opt/openfst
+COPY --from=openfst-builder /opt/openfst /opt/openfst
 ENV OPENFST_ROOT /opt/openfst
 
 ARG JOBS=4
 
+# Install runtime and build dependencies
 RUN apt-get update && \
     apt-get upgrade -y && \
-    apt-get -y install \
+    apt-get install -y --no-install-recommends \
     cmake \
     g++ \
-    libicu-dev
+    make \
+    ccache \
+    libicu-dev \
+    && rm -rf /var/lib/apt/lists/*
 
 RUN mkdir /fstalign
 COPY CMakeLists.txt /fstalign/CMakeLists.txt
@@ -24,10 +59,12 @@ COPY sample_data /fstalign/sample_data
 
 WORKDIR /fstalign
 
-RUN mkdir -p /fstalign/build && \
+RUN --mount=type=cache,target=/root/.ccache,sharing=locked  \
+    PATH=/usr/lib/ccache:${PATH} \
+    mkdir -p /fstalign/build && \
     cd /fstalign/build && \
     rm -rf * && \
-    cmake .. -DOPENFST_ROOT="${OPENFST_ROOT}" -DDYNAMIC_OPENFST=ON && \
+    cmake .. -DOPENFST_ROOT="${OPENFST_ROOT}" -DDYNAMIC_OPENFST=OFF && \
     make -j${JOBS} VERBOSE=1 && \
     mkdir -p /fstalign/bin && \
     cp /fstalign/build/fstalign /fstalign/bin && \

src/FstFileLoader.h

@@ -3,7 +3,8 @@
  * 
  * FstLoader for loading a serialized FST from disk.
  * 
- * Quinn McNamara (quinn@rev.com)
+ * Jp Robichaud (jp@rev.com)
+ * Quinn McNamara ()
  * 2020
  */
 

src/wer.h

@@ -3,7 +3,8 @@
  *
  * Collection of functions specific to the WER subcommand.
  *
- * Quinn McNamara (quinn@rev.com)
+ * Jp Robichaud (jp@rev.com)
+ * Quinn McNamara ()
  * 2021
  */
 #include "fstalign.h"